Re: [GENERAL] Upgraded to 9.3

2013-12-27 Thread Albe Laurenz
ERR ORR wrote: > This is just to say that I upgraded from V9.2.6 to V9.3.2. > > Same HW, same data, an astonishing performance increase. > Postgres is flying It lies in the nature of a support mailing list that most of the feedback you get is like "we upgraded, and this and that is much slower no

Re: [GENERAL] postgresql-common breaks logrotate

2013-12-27 Thread Magnus Hagander
On Fri, Dec 27, 2013 at 12:11 AM, Steve Crawford < scrawf...@pinpointresearch.com> wrote: > People who have attempted to install PGDG versions of PostgreSQL on recent > Ubuntu releases have run into the cascading problem of postgresql wanting > to destroy Ubuntu. > > Based on the packaging: > post

[GENERAL] ON_ERROR_EXIT and interactive mode (or, how to force interactive mode off)

2013-12-27 Thread Tim Kane
HI all, I have a fairly complex python script which calls a number of ‘psql’ sessions via a subprocess to perform bulk load operations. The bulk loads are generated on the fly, and are written directly to the stdin of the subprocess. I would like to rely on the ON_ERROR_STOP=1 functionality to f

[GENERAL] Why is wal_writer_delay limited to 10s?

2013-12-27 Thread Clemens Eisserer
Hi, Just to be curious, why is wal_writer_delay limited to 10s? I am using postgresql in an embedded environment where every 10s sensor values are logged and even with "synchronous_commit = off" and wal_writer_delay=1 this burns quite a lot of nand cycles. For me it wouldn't hurt loosing minut

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Joseph Kregloh
> > >> So how are you moving the 9.0 directory? > Just using a mv command like below: mv /usr/local/pgsql/data /usr/local/pgsql_90/data Then I recreated the symlinks in /pg_tbspc to point to the new directory path. What does a listing for that directory look like after the move? > [pgsql@postgr

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Adrian Klaver
On 12/27/2013 06:50 AM, Joseph Kregloh wrote: So how are you moving the 9.0 directory? Just using a mv command like below: mv /usr/local/pgsql/data /usr/local/pgsql_90/data Then I recreated the symlinks in /pg_tbspc to point to the new directory path. Ah, now I see the problem, I think

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Joseph Kregloh
> > > Postgres is going to /usr/local/pgsql/data/drupal_dbspace/ to look for > the 9.0 files instead of /usr/local/pgsql_90/data/drupal_dbspace/ and is > trying to copy them as 9.3 versions into the new default location which has > the same path. Since the new > /usr/local/pgsql/data/drupal_dbspac

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Adrian Klaver
On 12/27/2013 01:00 PM, Joseph Kregloh wrote: Postgres is going to /usr/local/pgsql/data/drupal___dbspace/ to look for the 9.0 files instead of /usr/local/pgsql_90/data/__drupal_dbspace/ and is trying to copy them as 9.3 versions into the new default location which has the sa

Re: [GENERAL] ON_ERROR_EXIT and interactive mode (or, how to force interactive mode off)

2013-12-27 Thread Chris Curvey
On Fri, Dec 27, 2013 at 7:30 AM, Tim Kane wrote: > HI all, > > I have a fairly complex python script which calls a number of ‘psql’ > sessions via a subprocess to perform bulk load operations. > The bulk loads are generated on the fly, and are written directly to the > stdin of the subprocess. >

Re: [GENERAL] Why is wal_writer_delay limited to 10s?

2013-12-27 Thread Gavin Flower
On 28/12/13 02:01, Clemens Eisserer wrote: Hi, Just to be curious, why is wal_writer_delay limited to 10s? I am using postgresql in an embedded environment where every 10s sensor values are logged and even with "synchronous_commit = off" and wal_writer_delay=1 this burns quite a lot of nand

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Joseph Kregloh
> > FYI, some testing showed that playing around with spclocation in > pg_tablespace is not recommended. Do you happen to have more information about this? Because it would actually solve all my problems by moving the user created tablespaces out of the /data directory. But I would like more info

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Adrian Klaver
On 12/27/2013 01:56 PM, Joseph Kregloh wrote: FYI, some testing showed that playing around with spclocation in pg_tablespace is not recommended. Do you happen to have more information about this? Because it would actually solve all my problems by moving the user created tablespaces out

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Adrian Klaver
On 12/27/2013 01:00 PM, Joseph Kregloh wrote: Postgres is going to /usr/local/pgsql/data/drupal___dbspace/ to look for the 9.0 files instead of /usr/local/pgsql_90/data/__drupal_dbspace/ and is trying to copy them as 9.3 versions into the new default location which has the sa

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Jeff Janes
On Friday, December 27, 2013, Joseph Kregloh wrote: > FYI, some testing showed that playing around with spclocation in >> pg_tablespace is not recommended. > > > Do you happen to have more information about this? Because it would > actually solve all my problems by moving the user created tablespa

Re: [GENERAL] pg_upgrade & tablespaces

2013-12-27 Thread Adrian Klaver
On 12/27/2013 02:52 PM, Jeff Janes wrote: On Friday, December 27, 2013, Joseph Kregloh wrote: FYI, some testing showed that playing around with spclocation in pg_tablespace is not recommended. Do you happen to have more information about this? Because it would actually